[FFmpeg-devel] [PATCH] vf_psnr: always calculate MSE over full pixel range.

[Paul B Mahol]

Dana 11. 7. 2015. 13:23 osoba "Ronald S. Bultje" <rsbultje at gmail.com>
napisala je:
>
> This makes the output compatible with that of pretty much any other
> tool that calculates PSNR.
> ---
>  libavfilter/vf_psnr.c | 31 ++++---------------------------
>  1 file changed, 4 insertions(+), 27 deletions(-)
>
> diff --git a/libavfilter/vf_psnr.c b/libavfilter/vf_psnr.c
> index 0c2c950..cc4546a 100644
> --- a/libavfilter/vf_psnr.c
> +++ b/libavfilter/vf_psnr.c
> @@ -260,33 +260,10 @@ static int config_input_ref(AVFilterLink *inlink)
>          return AVERROR(EINVAL);
>      }
>
> -    switch (inlink->format) {
> -    case AV_PIX_FMT_GRAY8:
> -    case AV_PIX_FMT_GRAY16:
> -    case AV_PIX_FMT_GBRP:
> -    case AV_PIX_FMT_GBRP9:
> -    case AV_PIX_FMT_GBRP10:
> -    case AV_PIX_FMT_GBRP12:
> -    case AV_PIX_FMT_GBRP14:
> -    case AV_PIX_FMT_GBRP16:
> -    case AV_PIX_FMT_GBRAP:
> -    case AV_PIX_FMT_GBRAP16:
> -    case AV_PIX_FMT_YUVJ411P:
> -    case AV_PIX_FMT_YUVJ420P:
> -    case AV_PIX_FMT_YUVJ422P:
> -    case AV_PIX_FMT_YUVJ440P:
> -    case AV_PIX_FMT_YUVJ444P:
> -        s->max[0] = (1 << (desc->comp[0].depth_minus1 + 1)) - 1;
> -        s->max[1] = (1 << (desc->comp[1].depth_minus1 + 1)) - 1;
> -        s->max[2] = (1 << (desc->comp[2].depth_minus1 + 1)) - 1;
> -        s->max[3] = (1 << (desc->comp[3].depth_minus1 + 1)) - 1;
> -        break;
> -    default:
> -        s->max[0] = 235 * (1 << (desc->comp[0].depth_minus1 - 7));
> -        s->max[1] = 240 * (1 << (desc->comp[1].depth_minus1 - 7));
> -        s->max[2] = 240 * (1 << (desc->comp[2].depth_minus1 - 7));
> -        s->max[3] = (1 << (desc->comp[3].depth_minus1 + 1)) - 1;
> -    }
> +    s->max[0] = (1 << (desc->comp[0].depth_minus1 + 1)) - 1;
> +    s->max[1] = (1 << (desc->comp[1].depth_minus1 + 1)) - 1;
> +    s->max[2] = (1 << (desc->comp[2].depth_minus1 + 1)) - 1;
> +    s->max[3] = (1 << (desc->comp[3].depth_minus1 + 1)) - 1;
>
>      s->is_rgb = ff_fill_rgba_map(s->rgba_map, inlink->format) >= 0;
>      s->comps[0] = s->is_rgb ? 'r' : 'y' ;
> --
> 2.1.2
>

LGTM

> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org
> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel